74LS76 and 74F112PC JK flip flops weird behavior

[Mohammed Hamed]

I tried to hook a simple circuit using a 74F112PC or 74LS76 (both get
the same behavior) dual negative edge triggered flip flop, but I get
weird behavior when both JK inputs are set to high. Whenever I made
the toggle manually by toggling the clock input, the flipping
(complementing) action doesn't always occur. This could be explained
easily by switch bouncing/debouncing. Although I don't know how much
bouncing exists in regular DIP switches. But the most puzzling thing
happens when I tie the Clock input to a regular square wave clock
(generated off a logic training board. The clock doesn't show any
glitches on oscilloscope output. However, sometimes the output doesn't
change, sometimes it follows the clock faithfully when it's supposed
to change on the negative edge, sometimes it doesn't work unless the
clock input pin is tied to an onboard LED, and sometimes it works well
when i touch the plastic coating of the wire between the clock and the
onboard LED and usually not all the LEDs. This weird behavior happened
with all the chips that i have for these 2 families, and it's
interesting that it's happening across families. I don't know what it
could be. The fact that touching the wire made me think it needs some
path to ground to lose extra charge (although i'm touching the
plastic). I tried to tie a large resistor across the clock input with
no much success. Do you have any ideas?

Thanks

 

[chris.welsh]

How is the circuit grounded? It sounds like there might be a floating
ground somewhere...

 

[CFoley1064]

Subject: 74LS76 and 74F112PC JK flip flops weird behavior

From: [email protected] (Mohammed Hamed)
Date: 2/26/2004 9:36 PM Central Standard Time
Message-id: <[email protected]>

I tried to hook a simple circuit using a 74F112PC or 74LS76 (both get
the same behavior) dual negative edge triggered flip flop, but I get
weird behavior when both JK inputs are set to high. Whenever I made
the toggle manually by toggling the clock input, the flipping
(complementing) action doesn't always occur. This could be explained
easily by switch bouncing/debouncing. Although I don't know how much
bouncing exists in regular DIP switches. But the most puzzling thing
happens when I tie the Clock input to a regular square wave clock
(generated off a logic training board. The clock doesn't show any
glitches on oscilloscope output. However, sometimes the output doesn't
change, sometimes it follows the clock faithfully when it's supposed
to change on the negative edge, sometimes it doesn't work unless the
clock input pin is tied to an onboard LED, and sometimes it works well
when i touch the plastic coating of the wire between the clock and the
onboard LED and usually not all the LEDs. This weird behavior happened
with all the chips that i have for these 2 families, and it's
interesting that it's happening across families. I don't know what it
could be. The fact that touching the wire made me think it needs some
path to ground to lose extra charge (although i'm touching the
plastic). I tried to tie a large resistor across the clock input with
no much success. Do you have any ideas?

Thanks

Youdefinitely have a problem with the clock. For the LS76, you need a single
transition with a fast rise time -- if you don't get it, you're toast. You
can't do that with a switch alone. All a resistor between the switch and the
input will do is slow the rise and fall times (an RC network with the
capacitance of the input).

You'll want a logic change on the NGT, so here's a trick that might work if
you've got a couple of CMOS NOR gates handy (view in fixed font or M$ Notepad):

Debounced Clock Switch Using Hardware NOR
VCC
|
.-. 47K
| | ___
| | 22K .-----|___|--------.
T '-' 10K | |
--- ___ | ___ | __ |
.----o o--|___|-o----|___|---o--|>=| __ |
| 10 Ohm | |1 |o-----|>=| | CLK'
| --- .--|__| |1 |o-o----o
=== --- 1uF | .--|__|
GND | | |
=== === ===
GND GND GND



If you've got CMOS inverters, just replace the NORs with inverters -- it will
work just as well. The circuit uses hysteresis to overcome the slow
transitions at the switch.

Good luck
Chris

 

[Mohammed Hamed]

[email protected] (CFoley1064) wrote in message

Youdefinitely have a problem with the clock. For the LS76, you need a single
transition with a fast rise time -- if you don't get it, you're toast. You
can't do that with a switch alone. All a resistor between the switch and the
input will do is slow the rise and fall times (an RC network with the
capacitance of the input).

You'll want a logic change on the NGT, so here's a trick that might work if
you've got a couple of CMOS NOR gates handy (view in fixed font or M$ Notepad):

Thanks for the response. 2 things here. First, The clock has rise and
fall time on the orders of nanoseconds as checked on the oscilloscope.
Second, does that explain the behavior when the output follows the
clock for long a time (many seconds sometimes). My guess that if these
were random glitches, I would get this behavior every once in a while.
It doesn't explain either when touching the wire makes it work, as i
don't believe this will make the rise time any faster.

 

[Roy McCammon]

I tried to hook a simple circuit using a 74F112PC or 74LS76 (both get
the same behavior) dual negative edge triggered flip flop, but I get
weird behavior when both JK inputs are set to high. Whenever I made
the toggle manually by toggling the clock input, the flipping
(complementing) action doesn't always occur. This could be explained
easily by switch bouncing/debouncing. Although I don't know how much
bouncing exists in regular DIP switches. But the most puzzling thing
happens when I tie the Clock input to a regular square wave clock
(generated off a logic training board. The clock doesn't show any
glitches on oscilloscope output. However, sometimes the output doesn't
change, sometimes it follows the clock faithfully when it's supposed
to change on the negative edge, sometimes it doesn't work unless the
clock input pin is tied to an onboard LED, and sometimes it works well
when i touch the plastic coating of the wire between the clock and the
onboard LED and usually not all the LEDs. This weird behavior happened
with all the chips that i have for these 2 families, and it's
interesting that it's happening across families. I don't know what it
could be. The fact that touching the wire made me think it needs some
path to ground to lose extra charge (although i'm touching the
plastic). I tried to tie a large resistor across the clock input with
no much success. Do you have any ideas?

I suspect ringing of the clock signal at the clock input of the flip
flop. It there is only one load on the clock, try a small (30-75) ohm
resister in series with the clock at the clock driver.

 

[Mohammed Hamed]

I suspect ringing of the clock signal at the clock input of the flip
flop. It there is only one load on the clock, try a small (30-75) ohm
resister in series with the clock at the clock driver.

I think the problem is with clock ringing, I drove the circuit from a
clean clock source and it works fine. I tried the small series
resistor but with no much success. Also I find that the instances in
which the flip flop is working have always the clock with no ringing
or very less. This happens when I connect an LED across the clock. The
LED is in series with 5k resistance and a darlington pair from a
ULN2803A chip.

My questions here are: If the ringing is not too severe to alter the
voltage levels, why does it cause the chip to malfunction? What I
understand is that the logic gates inside the flip flop have predefine
voltage levels with some noise margins that are most likely than the
ringing noise.

A 2nd question is :
Why the series combination of the 5k resistor and Darlington pair
across the clock input pin (i think the LED has no effect), why they
almost eliminate the ringing?

Thanks,